Unlimited-OCR
Developed by baidu
Unlimited-OCR is an open-source, high-performance document parsing model released by Baidu, designed to advance DeepSeek-OCR into the era of one-shot long-horizon parsing. It supports single-image, multi-page, and PDF document parsing with dual configurations: 'gundam' (crop mode) and 'base'. Fully integrated with high-efficiency inference engines like vLLM and SGLang, it delivers exceptional throughput and accuracy for enterprise-level document digitization.
- One-shot long-horizon parsing
- Multi-page PDF and multi-image stream parsing
- VLLM and SGLang integration
- Dual-mode (Gundam crop and Base) support
webdesktop